Research On Pedestrian Detection Before Vehicle And Distance Algorithm Based On Image Processing

In this paper,the pedestrian detection and the vehicle distance algorithm are the part of the Intelligent and Auxiliary Vehicle Driving System.Firstly,the methods and principles of pedestrian detection are summarized,and then I analyze particularly the image processing techniques and principles of pedestrian detection process.Pedestrian characteristics,the principle and process of feature selection.The algorithm of HOG(Histogram of Oriented Gradient)feature and SVM(Support Vector Machine)is used to identify the classification of pedestrians,according to the characteristics of the pedestrian edge,gradient the characteristics of pedestrians are extracted and analyzed.while building software platform,in the Visual Studio 2010 framework with OPenCV function and algorithm,we can come ture the static pedestrian detection.The distance between the pedestrian and the vehicle is calculated by the computer vision single head distance measurement geometric relationship.Finally,the distance between the vehicle and the vehicle calculated by the algorithm program is compared with the actual measured distance.The accuracy and error rate of the algorithm.Then,according to the algorithm of the design and development,the actual vehicle test is carried out.The accuracy of the pedestrian detection and the vehicle distance algorithm is verified by constantly comparing the distance between the vehicle and the vehicle distance and the algorithm.The experimental results are analyzed,Proposed improvement programs for future research to provide a reference.